Copula-Based Approach to Construct a Joint Probabilistic Model of Earthquakes and Strong Winds

Current structural design codes usually treat multiple hazards separately, and probabilistic backbones are rare for extreme hazard combinations, e.g., earthquake and strong wind, which may cause unforeseen damage to engineering structures exposed to multiple extreme hazards during their lifecycles. This study presents an innovative copula-based approach to construct the joint cumulative distribution function (JCDF) of the peak ground acceleration (PGA) and strong wind speed (V). Six commonly used Archimedean copulas are applied to bond the JCDF with the corresponding marginal cumulative distribution functions (MCDFs) of PGA and V. A total of 76 low-probability-high-consequence extreme events with a simultaneously occurring earthquake and strong wind are abstracted from data recorded from 1971-2017 in Dali Prefecture, China. The statistical analysis results show that the Frechet and truncated Weibull distributions are the optimal expressions for the marginal distributions of PGA and V, respectively, while the Joe Archimedean copula can yield good JCDF estimation. Monte Carlo simulation is employed to establish a target dependent multihazard database that can be used for the performance-based design of engineering structures against multiple natural hazards. A high-rise building is used to study the performance under the multihazard of an earthquake and strong wind. The results show that the maximum inter-story drift ratio of the building under multiple hazards increases by 14.4-21.3% compared with the structural response induced by an earthquake alone.